2024.2: More voice, more icons, more integrations, more... everything!

The longer term prospects for the Met Office integration are uncertain. Even if the Python incompatibility is fixed (and an upstream PR has been submitted, so awaiting the code owner to accept it), the current API it uses is I believe shutting down before the end of April 2024, and a new integration will need to be coded more or less from scratch. There’s more information in the GitHub discussion here.

I’ve moved to OpenWeatherMap in the mean time, and signed up to use its One Call API, which is free up to a limit which I’ve comfortably yet to exceed. Seems an adequate replacement.

2 Likes

Ever since the update, my Switchbot Blind tilts using the Switchbot integration has this problem where I can’t see the current status. They still work fine, but every time I control them I’ll get a message that the service call failed. And when I check the status, instead of telling me that it’s opened to a certain percentage or closed, it only says “Unknown”.

I have other devices using the Switchbot integration like the curtain and button pusher, but only the Blinds tilts are affected. I did ask around and one other person told me that they have this issue too. Is the team aware that this might be widespread yet.

Looks like there is an open issue:

Home Assistant Core 2024.2.0 Switchbot Blind Tilt bluetooth not reporting State · Issue #109947 · home-assistant/core (github.com)

1 Like

Okay, good to hear. Thanks.

time triggers only fire on the second, they don’t account for microseconds. So it is throwing the calc off. This is the most accurate I can get it scheduling the update ahead of time

- trigger:
    - platform: homeassistant
      event: start
    - platform: time
      at: sensor.next_swatch_time
  sensor:
    - name: Next SwatchTime
      unique_id: 8a98c032-e92b-4295-8b79-4fde63865e94
      device_class: timestamp
      state: >
        {% set rdt = utcnow() %}
        {% set inc = timedelta(seconds=86.4) %}
        {% set dt = rdt + timedelta(hours=1) %}
        {% set d = timedelta(hours=dt.hour, minutes=dt.minute, seconds=dt.second, microseconds=dt.microsecond) %}
        {% set c = d % inc %}
        {% set next = (rdt - c + inc).replace(microsecond=0) + timedelta(seconds=1) %}
        {% if next < rdt %}
          {{ (next + inc) | as_local }}
        {% else %}
          {{ next | as_local }}
        {% endif %}
    - name: Internet Time
      unique_id: f7e3e6b6-24c2-4ea3-9603-d466a1808b89
      state: >
        {% set time_bmt = utcnow() + timedelta(hours = 1) %}
        {% 
          set delta = timedelta(hours        = time_bmt.hour,
                                minutes      = time_bmt.minute,
                                seconds      = time_bmt.second,
                                microseconds = time_bmt.microsecond)
        %}
        {% set beat = (delta.total_seconds() * 10) | int // 864 %}
        {{ "@{0:03d}".format(beat) }}

If you just use the next template, it’ll have slightly more accuracy every 5th beat. All the other beats in between will have the same accuracy.

- trigger:
    - platform: time_pattern
      seconds: /1
  sensor:
    - name: Internet Time
      unique_id: f7e3e6b6-24c2-4ea3-9603-d466a1808b89
      state: >
        {% set time_bmt = utcnow() + timedelta(hours = 1) %}
        {% 
          set delta = timedelta(hours        = time_bmt.hour,
                                minutes      = time_bmt.minute,
                                seconds      = time_bmt.second,
                                microseconds = time_bmt.microsecond)
        %}
        {% set beat = (delta.total_seconds() * 10) | int // 864 %}
        {{ "@{0:03d}".format(beat) }}

Both versions will be at update at most 1 second off from the actual beat time update frequency.

1 Like

Ok hope that solves it, keep us posted! :pray:

I don´t have evohome, even I don´t know that it´s evohome

Yes that sounds logical as my reply was not on your message :sweat_smile:

In the right corner of the message there is a round icon with a reply arrow, when you click that you will see the message that is replied on :wink:

1 Like

Anyone else just lost all statistical data from the energy dashboard after update of today

First got message that unit of measurement wasn’t correct
And after reboot all data was gone

Reverted to before last release and works again

Still no matter info for my only matter device (a Nuki Smart Lock) after 2024.2.1 update :thinking:

I’m unsure if caused by HA 2024.2.0/1 and/or by the accompanying updates to OTBR/Matter addons, but my thread network is terribly unstable and less than one or two hours after a system reboot at least one (if not two or three) of my 25 HomeKit devices drop off the network, simply become “unavailable” until I reboot HA. This has been going on for the past few days.

I’m considering reverting to 2024.1, but am reluctant, as I haven’t yes reinstalled an older HA version from a full backup before and am fearing other issues.

I had these problems with versions matter versions +5.1.0, downgrading to 5.0.4 fixed the issues. Now I am testing 5.1.2.

There is a bug closed about this issue.

I appreciate the advice.
I solely had the first sentence originally. Then added the last three sentences to test/debug why it stopped working with just the “office to 75” but still worked with “office to 75 degrees” and posted them to show here, in this post, what I tried.
Not a big issue, I just have to add something after the “75”/{temp_degrees}

As a follow-up: I just checked (using only the first sentence as suggested) it again after today’s HA Core Update and Assist’s response is:
“Brightness Set” even though I don’t have “brightness” anywhere in my -intents:- nor my -intent_script:- code.

For the custom sentences users out there, this release improves range list number parsing.

lists:
  hours:
    range:
      from: 0
      to: 24
      digits: true# true by default; recognizes "1", "2"...
      words: true# true by default; recognizes "one", "two"...
      #words_language optional 2-letter code
      #words_ruleset optional, see below
lists:
  year:
    range:
      from: 0
      to: 3000
      digits: true
      words: true
      words_ruleset: "spellout-numbering-year"# recognizes "nineteen ninety-nine"

More values for this here.

Warning: with words: true, don’t go with large ranges… HA’s assistant still can’t do large math (in any reasonable timeframe, or without crashing), nor negative numbers :smile:

I use Firefox and this was happening as well. I opened it in Edge and it worked.

I restored a lot of Backups to revert to previous Versions without Bugs or Breaking Changes.

Mostly Full Backups, sometimes partial as well - at least this never had problems, everything went fine!

If u are unsure, just make a full backup of your actual System (and keep all backups somewhere else) as well!

Thanks. I found a partial backup from which I restored Matter Server addon 5.0.4 and my thread network seems to be a bit more stable, currently only 1/25 thread devices are unavailable. I’ll wait another day or two before doing a complete restore.

i think the same, i opened an issue for this.

The last Release Party introduced me to the -Developer Tools/Assist" feature. It seems there is a built in feature that overrides my custom sentences. Here is the output of that feature when I type “office to 73”:

office to 73
Language: English (en)

intent:
  name: HassLightSet
slots:
  area: office
  brightness: '73'
  name: all
details:
  area:
    name: area
    value: office
    text: office
  brightness:
    name: brightness
    value: 73
    text: '73'
  name:
    name: name
    value: all
    text: ''
targets: {}
match: true
sentence_template: '[<numeric_value_set>] <area> [to] <brightness>'
unmatched_slots: {}
source: builtin